I have been trying to build a preprocessing pipeline, but I am struggling a little to generate a list of the indexes for each column that is an object dtype. I have been able to get the names of each into an array using the following code:

categorical_features = [col for col in input.columns if input[col].dtype == 'object']

Is there an easy way to get the index of these columns, from the original input dataframe into a list, like this one that I built manually?

c = [1,3,4,5,6,7,8,9,10,11,12,14,15,16,17,18,19,20,21,22,23,24,25,28,29,
     30,31,38,39,40,41,42,43,44,45,50,51,55,56]

CodePudding user response：

I think you need select.dtypes and enumerate

df = pd.DataFrame({'A' : ['A', 'B', 'C'], 'B' : [1,2,3], 'C' : [1, '2', '3']})

print(df)

   A  B  C
0  A  1  1
1  B  2  2
2  C  3  3

idx_cols = [idx for idx, col in enumerate(df.select_dtypes('object').columns) ]

[0, 1]

CodePudding user response：

enumerate can help with that:

categorical_features_indexes = [i for i, col in enumerate(input.columns) if input[col].dtype == 'object']

CodePudding user response：

Use df.select_dtypes df.columns.get_indexer:

categorical_features = df.columns.get_indexer(df.select_dtypes('object').columns)

• df.select_dtypes returns a copy of df with only the columns that are of the specified dtype(s) (you can specify multiple, e.g. df.select_dtypes(['object', 'int'])).
• df.columns.get_indexer returns the indexes of the specified columns.